pem fuel cell testing and diagnosis || design and fabrication of pem fuel cell mea, single cell, and...

Chapter 2

Design and Fabricationof PEM Fuel Cell MEA,Single Cell, and Stack

PE

Cop

Chapter Outline

2.1. Introduction 4
4
2.2. MEA Design and Assembly 44

2.2.1. Gas Diffusion Layer

Design
44
2.2.2. Catalyst Layer Design
50
2.2.3. Proton Exchange

Membrane Design
57
2.2.4. MEA Assembly
58
2.3. Typical Examples for MEA

Fabrication 59

2.3.1. GDM Preparation
59
2.3.2. MPL Preparation
62
2.3.3. CL Fabrication
63
2.3.4. Membrane

Pretreatment
68
2.3.5. MEA Fabrication
68
2.4. Flow Field Design 70

2.4.1. Materials for a Flow

Field Plate
71
2.4.2. Flow Field Layout

(Channel Pattern)
71
2.4.3. Flow Channel

Parameters
72
M Fuel Cell Testing and Diagnosis. http://dx.doi.org/10.101

yright � 2013 Elsevier B.V. All rights reserved.

2.4.4. Flow Field Plate

Fabrication
72
2.5. Sealing Design 73

2.5.1. Sealing Material

Selection
73
2.5.2. Sealing Design and

Fabrication
73
2.6. Single Cell Design and

Assembly 74

2.6.1. Single Cell Hardware
74
2.6.2. Single Cell Assembly
74
2.7. Stack Design and Assembly 75

2.7.1. Hardware of a Fuel

Cell Stack
75
2.7.2. Heat Management in

a PEM Fuel Cell Stack
76
2.7.3. Fuel Cell Stack

Assembly
77
2.8. Chapter Summary 77

References 77

6/B978-0-444-53688-4.00002-4

43

http://dx.doi.org/10.1016/B978-0-444-53688-4.00002-4

44 PEM Fuel Cell Testing and Diagnosis

2.1. INTRODUCTION

The design and fabrication of proton exchange membrane (PEM) fuel cellcomponents, single cells, and stacks are two of the most important processes infuel cell technology development. In general, the design and assembly of a fuelcell can have a strong effect on its performance. Given the materials andcomponents used in the fuel cell, design and fabrication have to be optimizedwith respect to the corresponding fuel cell power output to achieve the bestperformance. To date, designs and assembly methods have been optimized andvalidated using fuel cell testing as well as real operation in various applicationsystems, such as portable power devices, stationary power generators, andautomobiles. The major challenges still hindering their commercialization arehigh cost and insufficient durability. The basic components of H2/air (O2) PEMfuel cells/stack have been briefly introduced in Chapter 1. In this chapter, thedesigns of the key components of PEM fuel cells and the resultant effects oncell performance will be discussed in detail, including the fabrication of thePEM fuel cell membrane electrode assembly (MEA), single cell, and stack.

2.2. MEA DESIGN AND ASSEMBLY

The MEA consists of an anodic electrode, PEM, and a cathodic electrode.Because the electrode reactions take place in the MEA, it is the heart of a H2/airPEM fuel cell. The components of an MEA include the anode gas diffusionmedium (A-GDM), anode microporous layer (A-MPL), anode catalyst layer(A-CL), PEM, cathode catalyst layer (C-CL), cathode microporous layer(C-MPL), and cathode gas diffusion medium (C-GDM), as shown in Fig. 2.1.A commonly used term is “gas diffusion layer” (GDL), which actually containsthe gas-diffusion-medium layer and the microporous layer. Each componentshown in Fig. 2.1 has specific characteristics and functions in fuel cell operationand performance. Therefore, they differ significantly in their design andfabrication. These topics will be addressed in the following sections.

2.2.1. Gas Diffusion Layer Design

The GDL is a key component in H2/air PEM fuel cells. As shown in Fig. 2.1, theGDL has a gas diffusion medium as a backing layer and a microporous layer asa sublayer. The A-CL or C-CL is coated on this MPL sublayer. The GDLperforms the following functions: (1) providing passages for gas diffusionthrough the GDL from the flow channel to the CL, (2) providing electronpathways from the CL to the flow field or vice versa, (3) retaining some wateron its surface to maintain proton conductivity through the PEM, (4) removingexcess water to prevent the CL and GDL from flooding, and (5) servingas a physical microporous support for the CL when it is applied onto the GDL.A GDL design must include the design of both the GDM and the MPL.

FIGURE 2.1 Schematic structure of the MEA, including A-GDM, A-MPL, A-CL, PEM, C-CL,

C-MPL, and C-GDM. (For color version of this figure, the reader is referred to the online version

of this book.)

45Chapter | 2 Design and Fabrication of PEM Fuel Cell

2.2.1.1. Gas Diffusion Medium

Carbon-fiber paper [1–4] and carbon-fiber cloth [5–10] with a thickness of 100–300 mm are widely used as the GDM for H2/air PEM fuel cells due to their highporosity (>70%) and good electronic conductivity. In Chapter 1, Fig. 1.6showed scanning electron micrographs of carbon fiber paper and cloth. Thetypical properties of these GDMs are listed in Table 2.1. Recently, metallicporous materials have also been explored as GDM candidates. For example,Zhang et al. [11] studied a 12.5-mm thick copper foil as a GDM, and the resultsshowed that the thinness and straight-pore feature of this material improvedwater management.

The hydrophobicity/hydrophilicity of a GDM is an important property forgas transportation. Although commercially available carbon paper and/orcarbon cloth are hydrophobic, they are usually pretreated using hydrophobicmaterials (such as fluorinated ethylene propylene (FEP) [13] and polytetra-fluoroethylene (PTFE) [14,15]), which seems to be a necessary step to furtherincrease their hydrophobicity and thereby prevent “water flooding.” The detailsof GDM pretreatment will be covered in Section 2.3.1.

2.2.1.2. Microporous Layer

TheMPL is a substrate layer applied onto the GDM, and it consists of carbon orgraphite powders with a hydrophobic material (such as PTFE) as a binder. Inthe industry, the MPL is also known as a carbon sublayer. The MPL thickness istypically optimized to suit the fuel cell operating conditions [16,17], while theMPL’s average pore diameter is <20 mm [18,19]. The primary purpose of the

TABLE 2.1 Typical Properties of Carbon Fiber Paper and Carbon Fiber

Cloth [12]

Method

Carbon-Fiber

Paper*

Woven

Fabric**

Thickness (mm) Calipers at 7 kPa 0.19 0.38

Areal weight (g m�2) Gravimetric 85 118

Density (g cm�3) At 7 kPa calculated 0.45 0.31

Resistance (through-plane, U cm2)

Two flat graphite blocksat 1.3 MPa

0.009y 0.005y

Bulk resistivity(through-plane. U cm)

Mercury contacts 0.08 Not available

Bulk resistivity(in-plane, U cm)

Four-point probe 0.0055z 0.009z

Gas permeability(through-plane, Darcys)

Gurley 4301permeometer

8$ 55$

Material description TorayTGP-H-060

Avcarb1071 HCB

* Reported by Toray (unless indicated otherwise).

** Reported by Ballard Material Systems (unless indicated otherwise).yMeasured at General Motors (GM), includes diffusion-medium bulk resistance and two contact

resistances (plate to diffusion media).zMeasured at GM. uncompressed, average of resistivity in machine, and crossmachine direction.$Measured at GM, uncompressed, see Eqn (2.12), 1 Darcy¼ 10�12 m2.


MPL is to improve water management [20,21] and redistribute the reactants, as(1) it can wick liquid water away from the CL to the GDL and thus facilitate gastransportation in the opposite direction [19]; (2) it changes the porosity of theGDM, preventing the catalyst ink from penetrating the GDM; and (3) it reducesthe contact resistance between the GDL and the adjacent CL and thus decreasesthe internal resistance of a PEM fuel cell.

Normally, the MPL is fabricated by thoroughly mixing carbon or graphitepowders, solvents (such as ethanol or isopropanol), and binders (such as PTFE)to form a paste that is then spread onto the GDM using the doctor bladetechnique [22], spraying [23,24], painting [25], rolling [24], or screen printing[24]. The GDM with MPL is afterward put into an oven and heated slowly to350 �C to evaporate the solvent and other organics, and to sinter the binder. Theproperties of an MPL thus formed can be changed by adjusting its composition(e.g. the loadings of binder and carbon) [13,18,26–28] and the properties of thecarbon or graphite powders [16,17,22,23,29,30].


2.2.1.3. Effect of PTFE Loading in GDL on Fuel Cell Performance

As a hydrophobic material, PTFE serves as a binder to maintain the integrity ofthe carbon particles in the MPL and provides suitable hydrophobicity to avoidwater flooding. The amount of PTFE in the GDL has a significant influence onfuel cell performance. Depending on the MEA structure and the fuel celloperating conditions, the optimized PTFE content reported in the literatureusually varies from 10 to 40% by weight [27,28,31–33]. For example, Lufranoet al. [32] reported that the best fuel cell performance could be achieved witha PTFE loading of 20 wt.% in a system operated at 70 �C, with 2.5- and 3.0-baroperating pressures for H2 and air, respectively. The same optimal PTFEloading was reported by Park et al. [27]. Giorgi et al. [28] investigated elec-trodes with different PTFE contents in the diffusion layer using Hg-instrusionporosimetry, scanning electron microscopy (SEM), and electrochemicaltechniques such as cyclic voltammetry, galvanostatic polarization, and elec-trochemical impedance spectroscopy. Their results indicated that the totalporosity and macroporosity of the GDL decreased with increasing PTFEcontent, as shown in Fig. 2.2 [28], resulting in decreased active catalystsurface area and gas permeability. At a high current density (in the masstransfer control region), both the mass transport rate and cell performanceincreased with decreasing PTFE content, due to the increased total porosity.The best fuel cell performance was obtained with the lowest PTFE content(i.e. 10 wt.%).

FIGURE 2.2 Total porosity, porosity (2–50 mm), and macroporosity of a GDL as a function of

the PTFE content [28].


2.2.1.4. Effects of MPL Carbon Loading on Fuel Cell Performance

Carbon loading in the MPL is directly related to the thickness, porosity, andelectrical conductivity of a layer. A thinner MPL may not properly perform itswater/gas management function. However, a thicker MPL means a longerpathway, which will increase the mass transport resistance. Therefore, celloperating conditions dictate the optimal carbon loading to obtain the best cellperformance. For example, the maximum fuel cell performance was achievedat 75 �C ambient pressure and 100% relative humidity (RH) with a carbon(acetylene black) loading of 0.5 mg cm�2 in the MPL [18]. A higher carbonloading creates a thicker MPL and consequently a longer path for gas diffu-sion. But if the carbon loading is too small, the MPL may be too thin toimprove water management. The optimal carbon loading depends on theproperties of carbon. For example, Jordon et al. [16] reported that the optimalloading for acetylene black carbon with 10 wt.% PTFE is 1.9 mg m�2.Generally, the optimal amount in the MPL increases with decreasing carbonsurface area [23].

Different carbon blacks have different characteristics that affect the MPLporosity, pore size distribution, and electrical conductivity. The typical carbonblacks used in MPLs are ketjen black, Vulcan XC-72�, acetylene black, andtheir composites [16,22,23,29,30,34]. Passalacqua et al. [23] compared theperformance of MPLs made with Vulcan XC-72�, Shawinigan acetylene black(SAB), Mogul L, and Asbury 850 graphite, each carbon having the differentspecific surface areas listed in Table 2.2. The porosimetric characteristics of theGDLs obtained using these MPLs are listed in Table 2.3. The best performancewas achieved using SAB, which has a high pore volume and a small averagepore size. This result is attributable to improved mass transport and watermanagement. Jordan et al. [16] also reported that a GDL with acetylene blackas the MPL carbon powder showed a better cell performance than one withVulcan XC-72�. The same conclusion was reported by Antolini et al. [29]. Onestudy [33] used carbon cloth rather than carbon paper as the GDM, andinvestigated the effect of MPL carbon powders on fuel cell performance.

TABLE 2.2 Materials Used for MPL Preparation [23]

Material Surface Area (m2 g�1)

Asbury graphite 850 13

SAB 70

Mogul L 140

Vulcan XC-72� 250

TABLE 2.3 Porosimetric Characteristics of GDLs with Different Carbon

Powders in Their MPLs. APR Is the Average Pore Radius; APRp and APRs

Are the Average Pore Radius for the Primary and Secondary Pores,

Respectively; and Vp and Vs Are the Specific Pore Volume for the Primary

and Secondary Pores, Respectively [23]

Carbon

in MPL

Pore Volume

(cm3 g�1)

APR

(mm)

Vp

(cm3 g�1)

Vs

(cm3 g�1)

APRp

(mm)

APRs

(mm)

Asbury 850 0.346 3.5 0.212 0.134 0.29 8.6

SAB 0.594 1.7 0.368 0.226 0.27 4.3

Mogul L 0.276 6.0 0.157 0.119 0.20 13.6

VulcanXC-72�

0.489 1.8 0.319 0.17 0.24 4.9


Unlike with carbon paper, both sides of the carbon cloth were coated witha carbon/PTFE mixture to form a gas diffusion half-layer (GDHL) on each side,then the CL was applied onto one of the GDHLs. The results suggested that fuelcell performance can be improved under high pressure by using cathodes withVulcan XC-72� carbon powder on the catalyst side and acetylene black on thegas side.

Wang et al. [22,30] investigated acetylene black (AB), Black Pearls� 2000(BP), and their composite in MPLs and proposed a bifunctional MPL with anAB–BP composite. The best cell performance, with a peak power density of0.91 W cm�2, was achieved at 80 �C and 0.2 MPa using an MPL with 10 wt.%BP and 90 wt.% AB and a total carbon loading of 0.5 mg m�2 on each side ofthe gas diffusion backing (GDB; TGPH-030 Toray carbon paper). This result isattributable to the pore size distributions and the wettability of the pore walls.BP has a larger surface area and the ability to adsorb water internally [35], sothe MPL using BP was more hydrophilic. Moreover, BP has the most micro-pores and macropores and the least mesopores, as shown in Fig. 2.3, which canlead to low gas transport and easy occupation by liquid water, thus limiting themass transport in the MPL with BP. Compared to BP, AB has fewer macroporesbut more mesopores, and these can provide more passages for gas transport inthe MPL. In addition, it has been reported [30] that the surface of AB is morehydrophobic than that of BP, so an MPL with AB provides more hydrophobicpores for gas transport. However, the hydrophobic pores in AB might hinderliquid water removal. On combining the advantages of AB and BP, it is foundthat a composite of the two carbon powders in a suitable ratio will retain thehydophobicity of AB and simultaneously provide more passages for gastransport. Experimental results [22,30] showed that the micropores for water

FIGURE 2.3 Pore size distributions in GDLs (including MPL on GDB (TGPH-030 Toray carbon

paper)) with three carbon powders in the MPL: BP, AB, and AB–BP composite (CC1, 80 wt.% AB

and 20 wt% BP with a loading of 0.5 mg cm�2 on each side of the GDB) [22]. (For color version

of this figure, the reader is referred to the online version of this book.)


flow were increased by adding a small number of BP in AB, while retaining thelatter’s hydrophobicity.

2.2.2. Catalyst Layer Design

The CL is where the electrochemical reactions occur, which makes it anotherkey component inside the MEA of PEM fuel cells. The CL is a uniform layerwith a thickness of 10–100 mm (usually <50 mm), composed of electrocatalystpowders, proton-conducting ionomer (e.g. Nafion�), and/or binder (e.g. PTFE).Almost all the important challenges in PEM fuel cell development, such as highcost and low durability, arise from the CLs because they are complex,heterogeneous, contain expensive Pt-based catalysts, and have low stability.The reactions in PEM fuel cells have three phases, involving the reactant gases(e.g. H2 or O2), proton conductive ionomer (e.g. Nafion�), and electronconductor (e.g. carbon-supported Pt catalyst). Therefore, when designing a CL,it is desirable to extend and maximize the three-phase reaction zone to optimizefuel cell performance.

The three-phase reaction boundary inside the CL is depicted in Fig. 2.4. Itcan be seen that every active reaction site must simultaneously possessa reactant gas, proton conductive ionomer, and electron conductor. The

FIGURE 2.4 Schematic of a three-phase

reaction boundary [36]. (For color version

of this figure, the reader is referred to the

online version of this book.)


passages for the transportation of the reactant gas, electrons, and protons mustbe tailored to the reaction zones. In addition, water is required to maintain theproton conductivity of the ionomer. However, water produced by the electro-chemical reactions in the CL must be removed, so passages for water transportin the reaction zones are also necessary.

Anode and cathode CLs must be designed to generate high rates for thedesired reactions and minimize the amount of expensive catalyst required toachieve the target performance. In addressing these requirements, the ideal CLsshould (i) maximize the active surface area per unit mass of the electrocatalysts,(ii) minimize the obstacles for reactant transport to the catalyst, (iii) enableproton transport to the exact required position, and (iv) facilitate water removal.These are also the main requirements in extending the three-phase reactionboundary. To meet these requirements, each material’s property specificationsshould be considered during designing. Some compromise between conflictingrequirements is also necessary. In addition, the CL structure should be carefullytailored by using materials that permit the proper interactions betweencomponents.

The important properties of a CL, such as electron and proton conductivi-ties, porosity, surface area, and catalytic activity, are determined by its struc-ture, fabrication method, and component properties. In the development ofPEM fuel cells to date, many kinds of CL fabrication methods have been used,which include the doctor blade technique, painting, printing, spraying, rolling,screening, and others. Some CL structures, such as the PTFE-bonded electrode,the Nafion�-bonded electrode, and the catalyst-coated membrane, have beenwell developed. The first generation of CLs, using PTFE-bonded Pt blackelectrocatalysts, exhibited excellent long-term performance but at a prohibi-tively high cost [37]. These conventional electrodes generally featured highplatinum loading, that is, 4 mg cm�2. One of the most significant improve-ments has been made by Raistrick [38], who fabricated a CL with dispersedPt/C, which was followed by painting/spraying a solubilized ionomer on itssurface. These electrodes used 0.4 mg cm�2 and demonstrated the sameperformance as did the first-generation electrodes with 4 mg cm�2 [38]. Furtherresearch has led to the lowering of Pt loading to 0.1–0.3 mg cm�2 by using thin-film methods [39–42], and even to 0.01–0.02 mg cm�2 with the sputtering-deposition method [43].


During PEM fuel cell development, two typical classes of CLs have beenexplored: hydrophobic and hydrophilic. These will be addressed in thefollowing sections.

2.2.2.1. Hydrophobic CL

PTFE-bonded hydrophobic CLs are commonly developed for H2/air PEM fuelcells. In this type of CLs, the catalyst particles (e.g. Pt/C) are thoroughlymixed with a certain amount of hydrophobic binder (such as PTFE, poly-vinylidene difluoride) to form a catalyst ink, then the ink is cast onto the GDL.To provide ionic transport to the catalyst site, PTFE-bonded CLs are generallyimpregnated with an ionomer (commonly Nafion�) by brushing or spraying,forming a typical gas diffusion electrode (GDE) with a hydrophobic CL. ThePTFE-bonded hydrophobic CL was a remarkable advance for PEM fuel celldevelopment. First, the substitution of platinum black with carbon-supportedplatinum decreased the platinum loading >10-fold while still achievinga similar performance. More importantly, proton transport was enhancedsignificantly by impregnating the CL with a proton-conducting material. Withexperimental progress, a PTFE content of 20–40 wt.% and a ratio of Nafion�

to carbon (in Pt/C catalyst) of 0.8–1.0 have proven to be the optimal param-eters for creating an efficient electrode. To date, the performance of suchelectrodes has been significantly improved, and mass manufacturing has beenachieved.

Figure 2.5 shows a schematic of the structure of a PTFE-bonded GDE. Theunique virtue of this electrode is that the gas transport limitation is significantlyreduced because the PTFE forms passages for gas transport. However, itsdisadvantages are obvious as well. The PTFE (especially with a high loading)may wrap around the catalyst particles and decrease both the electron

FIGURE 2.5 Schematic of a PTFE-bonded hydrophobic electrode [44]. (For color version of this

figure, the reader is referred to the online version of this book.)


conductivity and the catalyst utilization. In addition, application of Nafion� tothe electrode surface leads to asymmetric distribution because the sprayedNafion� cannot penetrate deeply into the CL. As a result, the catalyst particlesinside the CL may be inaccessible to the Nafion� ionomer, leading to a higherproton transport resistance and leaving some Pt inactive, as denoted by theblank circles in Fig. 2.5. According to estimates, the platinum utilization insuch electrodes is only 10–20% [45]. Finally, the MEAs assembled with suchGDEs are prone to delamination because the electrode and membrane swell todifferent degrees, which creates a discontinuity in the ion path and decreasesthe durability of the PEM fuel cells.

2.2.2.2. Hydrophilic CL

Unlike hydrophobic CLs, hydrophilic CLs use a hydrophilic perfluorosulfonateionomer (PFSI) such as Nafion� as a binder instead of PTFE. Hence, this kindof CL can be called an ionomer-bonded hydrophilic CL. During preparation,the catalyst powder (e.g. Pt/C), PFSI (e.g. Nafion�), and solvent (e.g. ethanol orisopropanol) are mixed thoroughly to form a uniform hydrophilic catalystink/paste that is then transferred to a GDL or a membrane. Hydrophilic CLs canbe classified into two groups, according to the transfer method: GDL-basedhydrophilic CL and catalyst coated membrane (CCM).

2.2.2.2.1. GDL-Based Hydrophilic CL

In the GDL-based hydrophilic CL, the hydrophilic catalyst ink/paste is coatedonto the GDL [31,46–48] with the same methods used in hydrophobic CLfabrication, such as brushing, spraying, and the doctor blade technique. Afterthe catalyst ink is spread, the electrode is first dried slowly at room temperatureand then dried at 80–135 �C for about 30 min. For example, Qi and Kaufman[47] produced a low Pt loading, high-performance electrode for PEM fuel cellsby casting catalyst ink made of E-TEK 20 wt.% Pt/C, Nafion�, and solventwater onto the GDL to form the catalyzed electrode. The best performance,with a peak power density of 0.72 W cm�2 under ambient pressure, was ach-ieved with a Nafion� loading of 30 wt.% and a Pt loading of 0.12 mg cm�2. Toimprove the fuel cell performance and catalyst utilization further, Qi andKaufman proposed various activation methods, such as steaming or boiling[49], high temperature and pressure operation [50,51], as well as H2 evolutionon the electrode [52]. In this CL structure, the catalyst particles came in contactwith the proton conductor (i.e. Nafion�). In this way, both electron and protontransfer were ensured, and Pt catalyst utilization was improved. However, thepassages for gas transport and water transfer can be limited by the lack ofhydrophobic agent (such as PTFE), so mass transfer could be an issue in the CLif “water flooding” occurs. To overcome this challenge, a thinner CL, such asa CCM electrode, may be required.


2.2.2.2.2. Catalyst Coated Membrane

To further increase the ionic connection between the membrane and the CL,and improve the mass transfer in a hydrophilic CL, a thin-film CL (TFCL) wasproposed by Wilson and Gottesfeld [39,41]. This TFCL involves casting thecatalyst ink onto the membrane directly rather than onto the GDL to forma CCM. Wilson and Gottesfeld [41,42] suggested a decal transfer method forfabricating the ionomer-bonded hydrophilic CL. This process includes two keysteps: coating the catalyst ink onto a blank substrate film and then transferringthe coat onto the proton conductive membrane (e.g. Nafion� membrane).

Besides Nafion� ionomer, cation (such as tetrabutylammonium ion (TBAþ)or Naþ) exchange ionomers and membranes, which have a higher glass tran-sition temperature, are often alternatively adopted in catalyst ink preparationand decal transfer [37,42,53,54]. By using this ionomer, one can increase thetransferring temperature to as high as 160–210 �C without any structuraldamage. The high temperature facilitates effective contact between the ionomerand the catalyst particles and forms a more intimate membrane/CL interface. Itcan also introduce a robust, pseudocrystalline structure to the ionomer inthe CL. After the decal transfer process, the catalyzed membrane assembly isconverted to the Hþ form by lightly boiling it in diluted H2SO4 and rinsing it indeionized water before incorporating it into the MEA for testing. This proce-dure is depicted in Fig. 2.6 as the “conventional decal method.”

Recently, a modified decal transfer technique for CCM fabrication wasreported [55]. In this method, a colloidal catalyst ink was used, as described byUchida [56]. First, the ink was coated onto a Teflon� substrate. After drying, theCLwas transferred to a Hþ formmembrane (e.g. Nafion� 112 membrane) by hotpressing at 120–135 �C. Finally, the Teflon� substrate was peeled off the CCM.

Conventional Decal Method Improved Decal Method

Solution ink(catalyst + ionomer+ solvent + NaOH)

Colloidal ink(catalyst + ionomer

+ solvent )

Hot press (transfer)temp. 180 °C

Hot press (transfer)temp 120-135 °C

Substrate peel off(Na+ form MEA)

Substrate peel off(H+ form MEA)

NaOHtreatment

H2SO4 treatment(H+ form MEA)

≥

FIGURE 2.6 Schematic flowcharts of the conventional decal method and the improved decal

method for making a CCM [55]. (For color version of this figure, the reader is referred to the

online version of this book.)


This procedure is depicted in Fig. 2.6 as the “improved decal method.” This is thesimpler of the two methods. In addition, the fuel cell testing results [55] indicatedthat the MEA made by the improved decal method yielded a better fuel cellperformance. The authors attributed the superior cell performance to the higherporosity of the agglomerates in the MEA, which facilitated mass transport.

The CCM technology has been well developed in recent years [57–60] dueto the advantages of having (i) a tight contact between the CL and themembrane to achieve low interfacial resistance, (ii) a thin CL with low masstransfer resistance, and (iii) good contact among the electrode components.Note that during the preparation of CCM catalyst ink, alcohol or isopropanol isgenerally used as the solvent rather than glycerol, which is used in theconventional decal method. In a CCM, catalyst utilization can also be improved,with the Pt loading reduced to levels as low as 0.07 mg cm�2 [39] and even0.02 mg cm�2 [59], yet yielding a highly satisfactory fuel cell performance.

Regarding the ionomer (PFSI) content in a hydrophilic CL, the optimalamount and distribution of the ionomer in the CL is a tradeoff among threerequirements: (i) maximum contact between the ionomer and the Pt particles toguarantee proton transport, (ii) minimal electron resistance, and (iii) minimalgas transport resistance. Normally, gas transport can be affected by bothdecreased porosity due to the presence of a solid ionomer and liquid wateraccumulation due to the hydrophilicity of the CL. When carbon-supportedplatinum (Pt/C) is used as the catalyst, the carbon particles have a much largersurface area than the Pt particles, so only if the carbon surface is covered by theionomer can contact between the ionomer and the Pt particles be ensured. Thisindicates that the ratio between the ionomer and the carbon in the CL is quiteimportant for achieving high performance. The suggested ratio of ionomer tocarbon (I:C) is about 0.8:1.0, which is calculated based on the assumption thatthe ionomer forms a thin layer (~1 nm) on the carbon surface.

2.2.2.3. Partially Pyrolyzed Nafion�-Ionomer-Bonded Electrode

As discussed above, in the structure of a conventional hydrophobic CL, thehydrophobic agent (e.g. PTFE) is normally used during the catalyst ink prep-aration, and a certain amount of proton conductor (ionomer, e.g. Nafion�) issprayed onto the CL surface. However, because the sprayed Nafion� solutionmay not effectively penetrate the interior of the CL, the three-phase reactionzone will not be extended sufficiently. As a result, the contact between thecatalyst particles and the proton conductor (e.g. Nafion�) might not be verytight, which leads to low Pt catalyst use. Conversely, in the structure ofa Nafion�-bonded hydrophilic CL, the catalyst particles and Nafion� ionomerhave good contact because they are mixed during the catalyst ink preparation.The passages for both electron and proton transfer are guaranteed, and Ptutilization is improved as well. However, there are too few passages in the CLfor the reactant gas and water, due to the lack of a hydrophobic agent.

Pt/C Nafion Alcohol

Slurry

Electrode

Ball milling

Electrode precursor 1

Electrode precursor 2

Spary Nafion

Paste to GDL

Baking at340°C280

FIGURE 2.7 Schematic flowchart for the preparation of partially pyrolyzed Nafion�-ionomer-

bonded electrode. (For color version of this figure, the reader is referred to the online version of

this book.)


To utilize the advantages of PTFE-bonded hydrophobic and Nafion�-bonded hydrophilic CLs, a novel electrode structure was proposed [56]. Inthis new structure, the catalyst slurry was made by mixing catalyst powders,Nafion� ionomer as the binder, and alcohol as the solvent. The preparedcatalyst slurry was then coated onto the GDL to obtain the electrodeprecursor, which was baked at 280–340 �C under nitrogen to pyrolyze someof the Nafion� ionomer. Part of the ionomer in the CL was partially pyro-lyzed during baking by controlling the temperature and time. Nafion� hastwo functional groups in its molecular structure: a sulfonic acid group anda fluorinated carbon chain. During pyrolysis at approximately 280 �C, part ofthe Nafion� ionomer will lose its sulfonic acid group, leaving the fluorinatedcarbon chain. The left carbon chain has properties similar to PTFE, and canserve as the hydrophobic agent and bonder in a CL, as PTFE does in a PTFE-bonded hydrophobic CL. However, the unpyrolyzed Nafion� ionomer in theCL retains its original structure, with the sulfonic acid group attached to thefluorinated carbon chain, and can therefore be the bonder and protonconductor, as Nafion� ionomer is in a Nafion�-bonded hydrophilic CL.Consequently, after baking, a certain amount of Nafion� was sprayed ontothe electrode surface to further extend the three-phase zone. The processesfor this electrode preparation are depicted in Fig. 2.7. In this partiallypyrolyzed Nafion�-ionomer-bonded electrode, the hydrophobic and

FIGURE 2.8 PEM fuel cell performances of MEAs with different electrodes. Nafion� 1035

membrane; cell temperature: 80�; backpressure: 0.2 MPa; humidity temperatures for H2 and air:

90� and 85�; stochiometries of H2 and air: 1.5 and 2.5 [56].


hydrophilic structures can be distributed uniformly in the CL. The transportof electrons, protons, reactant gases, and water can be facilitated. In addition,the three-phase reaction zone can be effectively extended. Figure 2.8 presentsa typical result, showing improved fuel cell performance and Pt utilization[56]. It is worth pointing out that the Nafion ionomer loading, bakingtemperature, and baking time are three key parameters in achieving anelectrode with high fuel cell performance [56].

2.2.3. Proton Exchange Membrane Design

As shown in Fig. 2.1, the PEM is another key component in the MEAs of H2/air PEM fuel cells, serving not only as a solid electrolyte but also as a sepa-rator to prevent direct contact between the anode and cathode compartments.The most practically and extensively used PEM in H2/air PEM fuel cells is theperfluorosulfonic acid (PFSA) membrane, such as Nafion� membrane. Asshown in Chapter 1, Fig. 1.7, the molecular structure of Nafion� consists ofthree groups: the tetrafluoroethylene (Teflon)-like backbone; the sulfonateacid group; and the –O–CF2–CF–O–CF2–CF2– side chain, which connects thebackbone and the sulfonate group. So far, PFSA membranes are considered tobe the best and have been widely used in H2/air PEM fuel cells because oftheir high proton conductivities and good stabilities in both oxidative andreductive environments. However, they also have disadvantages, such as highcost, high degradation rate at high temperatures (>80 �C), and dependence ofproton conductivity on the membrane’s water content. Thus, the developmentof new membrane materials continues to be a hot topic for research anddevelopment in PEM fuel cell technology. The requirements for PEM


materials are (1) low cost; (2) strong mechanical stability; (3) strong chemicalstability; (4) strong electrochemical stability in both oxidative and reductiveenvironments; (5) high proton conductivity in the operating temperature rangeof fuel cells; (6) insensitivity of proton conductivity to water content; and(7) low gas permeability.

In recent years, many kinds of materials have been developed to synthesizeproton-conducting membranes for H2/air PEM fuel cells, and some haveexhibited promising performance as potential candidates to replace PFSAmembranes. The major membranes are (1) fluorinated membrane, (2) partiallyfluorinated membrane, (3) nonfluorinated (including hydrocarbon) membrane,and (4) nonfluorinated composite membrane. Among these, the hydrocarbonmembrane is considered a promising alternative due to its low cost comparedwith PFSA membranes [61].

Several literature reviews [62–69] provide more detailed information aboutfuel cell membranes and their design and fabrication.

2.2.4. MEA Assembly

As a key component in PEM fuel cells, the MEA consists of the anode, PEM,and cathode. Its assembly mainly involves a hot-pressing process to makea “sandwich” with the membrane in the middle and the anode and cathode oneither side. Both the anode and the cathode should contain a GDM, MPL(carbon sublayer), and CL. When the MEA is being hot pressed, the CL mustface one side of the membrane. Before MEA assembly, pretreatment of themembrane is usually required to remove possible impurities and to completelyprotonate the membrane.

Hot pressing is an effective and simple way for assembling electrodes andPEMs to achieve good interfacial contacts between them. Hot-pressingconditions, such as temperature, pressure, and pressing duration, influence theperformance and durability of the resulting MEA [70–74]. A study [72] showedthat the combination of temperature, pressure, and time should be optimized toachieve a high-performance MEA.

Apparently, the temperature plays a major role in this optimization. Fora Nafion�-based membrane, the hot-pressing temperature is normally limitedby its glass transition temperature (Tg, ~128 C). At a temperature lower than Tg,the Nafion� resin in both the CL and the membrane will not melt and can resultin poor ionomeric contact between the CL and the membrane, which leads tolow catalyst utilization and higher ionic resistance. In contrast, a temperaturemuch higher than Tg may lead to a loss in the water retention properties ofNafion�, and acidic group degradation in the ionomer. Therefore, there is anoptimal temperature for the hot-pressing process [70,73,74]. Under this optimaltemperature, the CL and membrane combine most effectively to providea maximum electrochemical area at the interface, leading to the highest catalystutilization and fuel cell performance [70].


The hot-pressing pressure is related to the mechanical strength, porosity,and thickness of the electrode. Normally, this porosity decreases withincreasing pressure, which can restrict the mass transport of the gas. Moreover,the carbon fibers are prone to be crushed under high pressure. However, theelectrode thickness can be decreased under a high hot-pressing pressure, whichshortens the mass transportation pathway. A study [72] showed that a lower hot-pressing pressure could result in a better fuel cell performance than a higherhot-pressing pressure.

Hot-pressing time is another important parameter that affects the contactbetween the membrane and the electrode, as well as the electrode porosity. It isrecognized that with an increase in hot-pressing time, the ionic conductivityand the three-phase reaction area in the CL can be first increased and thendecreased, and the electrode porosity can also be decreased. Liang et al. [71]used direct methanol fuel cells to investigate the durability of MEAs assembledunder different hot-pressing conditions, and found that a longer hot-pressingtime could induce significantly improved MEA durability without sacrificingcell performance, because of a stronger interfacial binding between the CL andthe membrane, which suppresses their delamination. Although the optimalhot-pressing conditions for PEM fuel cells are slightly different due to thedifferences in the materials and structures of the electrode and membrane,hot-pressing of PEM fuel cell MEAs is usually conducted at 120–160 �C and2000–35,000 kPa pressure for 30–300 s.

Typically, the anode CL and cathode CL are applied onto their respectiveGDLs to form the anode and cathode GDEs, then the anode and cathode GDEsare hot pressed with the membrane in the middle to form a sandwich-likestructure. In recent years, there have been significant developments in theCCM, which is a typical three-layer MEA. For the CCM, however, the hot-pressing process joining the membrane and the anode and cathode GDLs isunnecessary. Usually, the GDLs are simply pressed together with the CCMusing blade pressure, during the fuel cell assembly process.

2.3. TYPICAL EXAMPLES FOR MEA FABRICATION

To provide a better understand of MEA fabrication, this section offers sometypical examples, including GDM, MPL, and CL preparation, membranepretreatment, and MEA assembly.

2.3.1. GDM Preparation

As described in Section 2.2.1.1, the GDM is the backing layer that supports theMPL and CL. The most commonly used materials for the GDM are carbon-based sheets, such as carbon paper and carbon cloth. Carbon paper is presentlymore widely used because it is relatively inexpensive, and MPLs and CLs are


easier to make on it than on carbon cloth. Section 2.3 on MEA fabrication thusdescribes the techniques that use carbon paper.

Although as-received carbon paper is to some extent hydrophobic, it is stillnecessary to increase its hydrophobicity by pretreating it before use in MEAfabrication, to prevent water flooding. This pretreatment method is known as“carbon paper wet proofing,” and the procedure is as follows: (1) dip the GDMinto an aqueous solution or suspension of a hydrophobic agent such as PTFE orFEP; assuming PTFE is used, the concentration can be varied from 1 to10 wt.%; (2) remove the GDM and eliminate the excess solution or suspension;(3) dry the PTFE-impregnated GDM in an oven; (4) repeat the above three stepsuntil the expected PTFE content in the GDM is achieved (usually 5–30 wt.%);(5) bake the impregnated GDM in an oven at 350 �C to remove the solvent andsurfactants contained in the PTFE suspension, to sinter the PTFE particles, andto fix the PTFE to the GDM surface. The desired PTFE content can easily bereached by adjusting the concentration of the PTFE suspension. But a lowerPTFE concentration and more dipping times are helpful in achieving a uniformPTFE distribution in the GDM. Figure 2.9 shows an SEM image of carbonpaper with 20% PTFE [2], wherein it is evident that the PTFE uniformly coversthe carbon paper surface.

Figure 2.10 shows SEM images of (a) carbon paper wet proofed with 20%FEP and (b) original carbon paper. Evidently, the FEP dispersed evenly on thesurface of the carbon paper. Figure 2.10 also shows that some pores on thecarbon paper surface were covered by FEP after wet proofing. These resultssuggest that this method can increase the hydrophobicity of the carbon paper,which can be characterized by measuring its contact angle.

FIGURE 2.9 SEM image of carbon paper (Toray, TGPH-090) pretreated with 20% PTFE [2].

(a)

(b)

FIGURE 2.10 Comparison of surface SEM images of (a) Carbon paper (Toray, TGPH-090,

E-TEK) impregnated with 20 wt.% FEP hydrophobic polymer, to (b) Untreated carbon paper [13].


For the wet-proofing treatment of carbon papers, the PTFE distributionthrough the thickness of the GDM is sensitive to the drying method. As shownin Fig. 2.11 [12], a fast drying method results in more concentrated PTFE on theexposed GDM surfaces. However, a slow drying method tends to form a moreuniform distribution of PTFE through the bulk of the GDM.

Note that the PTFE content or loading in the GDM must be optimized withrespect to the fuel cell performance. Too much PTFE in the GDL can decrease

FIGURE 2.11 Cross-sectional fluorine maps across carbon-fiber paper (Toray TGP-H-060).

PTFE distribution through the paper depends heavily on drying conditions [12].


its conductivity, and too little PTFE will lead to insufficient hydrophobicity forsmooth gas transportation.

2.3.2. MPL Preparation

As described in Section 2.2.1.2, the MPL is a sublayer of carbon powders madeon the GDB that can change the GDB porosity and support the CL. The typicalprocedure for MPL preparation is as follows:

(1) In a suitable ratio, ultrasonically mix the carbon powder (e.g. AB, BP, orVulcan X-72� carbon), the hydrophobic agent (e.g. PTFE), and the solventto create a uniform ink/paste.

(2) Coat the above ink/paste onto one or both sides of the GDB by brushing,spraying, spreading, doctor blade, screen printing, or by using othertechniques.

(3) Bake the carbon-coated GDM in an oven with nitrogen flow at 240 �C for30 min, followed by another 40 min at 350 �C; the MPL will then beformed on the GDB.

It must be noted that the MPL can be applied on one or both sides of the GDB,depending on the GDB material. For example, the MPL is usually created onboth sides of an SGL carbon paper but on only one side of a Toray TGPH carbonpaper. The loadings of carbon powder and PTFE can significantly affect theGDL performance. Hence, these loadings need to be optimized according to theMEA structure and operating conditions [18,27], as discussed in Section 2.2.1.2.The category of carbon powder used in the MPL also significantly affects theperformance of the GDL and, consequently, that of the fuel cell [22,30,75].

(a) (b)

(c)

FIGURE 2.12 SEM micrographs of cathode GDL surfaces with MPLs prepared using different

carbon powders: (a) Acetylene black (AB), (b) Black Pearls 2000 (BP), and (c) Composite carbon

with 90% AB and 10% BP [30].


Figure 2.12 shows SEM micrographs of the GDL surfaces with MPLsprepared using different carbon powders. It can be seen that the MPL with ABpossesses rich pores and a uniform surface, whereas the MPL with BP is denseand has only large cracks. These features can be attributed to the properties ofAB and BP carbon powders, which have been addressed in Section 2.2.1.2. Asshown in Fig. 2.12, an MPL with composite carbon powders (CC) presentsa more uniform surface with finer pores. The fuel cell testing results indicatethat the best performance is achieved using an MPL with CC, as shown inFig. 2.13. This can be explained by the more functional pore structure formedon the MPL when CC facilitates the mass transport of gas and water [22,30].

2.3.3. CL Fabrication

2.3.3.1. PTFE-Bonded Hydrophobic CLs

The PTFE-bonded hydrophobic CL is one of the classic CLs. The preparationprocess for this kind of CL is similar to the MPL preparation process and can bedescribed as follows [76–78]:

(1) Mix the catalyst powder (e.g. Pt/C), hydrophobic bonder (PTFE emulsion),and solvent (e.g. ethanol or isopropanol) ultrasonically to form a catalystink/paste;

FIGURE 2.13 H2/air PEM fuel cell performance of electrodes with MPLs prepared using

different carbon powders. Cell temperature: 80 �C; humidifier temperatures for anode and cathode:

90� and 85�, respectively; operating pressure: 0.2 MPa [30].


(2) Spread the catalyst ink/paste onto a GDL (with the MPL on wet-proofedcarbon paper) using the doctor blade technique or by spraying, painting,screen printing, etc.;

(3) Dry the precursor prepared in the previous step at 240 �C for 30 min underan inert gas atmosphere to remove the solvent and the surfactants containedin the PTFE emulsion;

(4) Bake the precursor at 350 �C for 30–40 min to sinter the PTFE and therebyhydrophobilize the CL;

(5) Spray a certain amount (e.g. 0.5 mg cm�2) of ionomer (e.g. Nafion) solu-tion onto the CL surface to form ionic pathways and increase the three-phase reaction zone;

(6) Dry the as-prepared electrode at room temperature to evaporate the solventcontained in the ionomer, and finally obtain the PTFE-bonded hydrophobicelectrode.

In this PTFE-bonded hydrophobic electrode, the PTFE content significantlyaffects the fuel cell performance. The passages for gas and water transport aretailored by introducing the PTFE during the catalyst ink/paste preparation stage.However, these proton conductors are not enough because the impregnatingNafion, located on the CL surface, cannot deeply penetrate the electrode. Thus,the three-phase reaction zone is not extended efficiently, leading to lowPt utilization.


2.3.3.2. Ionomer-Bonded Hydrophilic CLs

Unlike a PTFE-bonded hydrophobic CL, in an ionomer-bonded hydrophilicCL, the proton-conducting ionomer (e.g. Nafion) is used as the bonder. Theclassic preparation process for this CL is as follows [46,47,79,80]:

(1) Prepare the catalyst ink by thoroughly mixing the catalyst (e.g. Pt/C), ion-omer (e.g. Nafion), and solvent (e.g. alcohol, usually isopropanol);

(2) Spray the catalyst ink onto the GDL (with the MPL on wet-proofed carbonpaper);

(3) Dry the electrode first at room temperature, then at 70–135 �C for 30 min,to obtain an electrode with an ionomer-bonded hydrophilic CL on the GDL.

Figure 2.17 shows the degradation of an MEA made by applying a hydrophiliccatalyst ink to a GDL (i.e. using the conventional method). This MEA exhibitedgood durability.

Aside from being sprayed on the GDL, the catalyst ink can also be appliedto the membrane to make a Nafion-bonded hydrophilic CL. To efficientlyextend the three-phase reaction zone and reduce the Pt loading, Wilson et al.[37,39,41] developed a thin-film electrode using Nafion ionomer as the bonder.Their preparation process uses a decal method, the details of which are asfollows [39]:

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Ce

ll V

olta

ge

(V

)

Current Density (A·cm-2

)

Air

OXYGEN

Tcell/Thumidifiers = 80/105 oC Anode/Cathode = 3/5 atm

FIGURE 2.14 Air and oxygen fuel cell polarization curves for directly catalyzed developmental

Dow membrane with a catalyst loading of 0.13 mg Pt cm�2 [39].

FIGURE 2.15 A schematic representation of the procedure for MEA fabrication using the decal

transfer process [83]. (For color version of this figure, the reader is referred to the online version of

this book.)


(1) Prepare a uniform hydrophilic ink composed of Nafion solution (in Naþ

form), Pt/C catalyst, and solvent (e.g. glycerol). The weight ratio of Pt/Ccatalyst to Nafion is typically between 5:2 and 3:1.

(2) Paint this catalyst ink onto a dry Nafion membrane (in Naþ form) to forma CL, then bake the CL at approximately 160–190 �C to dry the ink.

(3) Cast the catalyst ink onto the reverse side of the Nafion membrane using thesame process.

(4) Rehydrate and ion exchange the membrane into the Hþ form by immersingthe catalyzed membrane into a slightly boiling 0.1 M sulfuric acid solutionfor 2 h.

(5) Rinse the catalyzed membrane sufficiently and air dry it.

The result is a thin-film, Nafion�-bonded hydrophilic electrode in which thecatalyst and ionomer are thoroughly mixed, but which lacks the passage for gasand water transport because it has no hydrophobic agent. So, this electrode isusually made very thin (5–10 mm) to avoid “water flooding” Fig. 2.14 shows thefuel cell performance of a thin-film electrode prepared by the above method. Itcan seen that the electrode exhibited a good performance with a Pt loading ofjust 0.13 mg cm�2.

Aside from the above process, a decal transfer method has also beendeveloped to make a thin-film electrode [81–83]. The catalyst ink is first coatedonto a decal substrate (such as PTFE film or Kapton� film) by spraying or byusing the doctor blade technique. The catalyst ink is then transferred to aNafion� membrane by a hot-pressing process to form a catalyzed membrane.This decal transfer method is presented schematically in Fig. 2.15 [83].

FIGURE 2.16 Polarization curves of MEAs made by the decal transfer method, with different

Nafion contents (N10, N20, N30, and N40), and at various RHs on the cathode, from 25 to 95%.

Cell temperature: 65 �C; stoichiometries of H2 and air: 1.5 and 2.0, respectively; anode RH was

fixed at 80% [82].


Figure 2.16 shows the polarization curves of MEAs made by this decal transfermethod with different Nafion� contents (N10, N20, N30, and N40), at RHsfrom 25 to 95% on the cathode. It can be seen from Fig. 2.16 that the Nafion�

content plays an important role in determining the fuel cell performance, andthe optimal Nafion� content is related to the operating conditions.

The CCM has been well developed in recent years [79,84,85]. During CCMpreparation, the catalyst, Nafion� ionomer, and solvent are mixed to forma uniform catalyst ink, which is then directly sprayed onto the membrane. In theCCM, the contact between the CL and the membrane is tight, so the CL doesnot tend to be delaminated during long-term fuel cell operation, suggestinggood MEA durability. As shown in Fig. 2.17, the CCM in study [79] exhibitedvery good durability, with a low fuel cell degradation rate during 1000 h of fuelcell testing.

FIGURE 2.17 Effects of MEA fabrication method on voltage degradation in single cells oper-

ated at 600 mA cm�2. Cell temperature: 80 �C; stoichiometries of H2 and air: 1.5 and 3.0, with

RHs of 100% and 55%, respectively [79].


2.3.4. Membrane Pretreatment

The PEM is the heart of an MEA, and it not only separates the anode fromthe other compartments but it also conducts the protons produced at theCL/membrane interface. The as-received membrane may contain impurities ormay not be fully protonated; either factor will affect the membrane’s perfor-mance and thus eventually influence the fuel cell’s performance. It is thereforenecessary to pretreat membranes before they are used in MEAs.

The typical membrane pretreatment process includes the following [72]: (1)boiling the membrane at 60–80 �C in a dilute H2O2 solution (H2O2 concen-tration of ~3–5%) to remove the organic and inorganic impurities contained inthe membrane; (2) boiling the membrane at 60–80 �C in an H2SO4 aqueoussolution (a 0.5 M H2SO4 solution is usually employed) to protonate themembrane; and finally, (3) twice rinsing the membrane at a high temperature,such as 70 �C, with deionized water. The resulting pretreated membrane can bestored for future use.

2.3.5. MEA Fabrication

Hot pressing is a widely used method to prepare conventional hydrophobic andhydrophilic electrodes for MEAs, and descriptions of the process can be found


elsewhere [71–73,76]. Hot-pressing parameters, such as temperature, pressure,and duration, significantly affect MEA performance, as shown in Fig. 2.18. Ifa CCM is used, it is usually put between the anode and cathode GDLs, whichare then pressed together by blade pressure during fuel cell assembly. Hotpressing of a CCM and GDLs is not necessary because the CL has good contactwith the membrane. Indeed, hot pressing reduces the porosity of GDLs andleads to decreased durability, as shown in Fig. 2.17.

C1 : P 500 psi. t 2 min. T 100°C

C2 : P 500 psi. t 5 min. T 100°C

1.2

1.0

0.8

0.6

0.4

0.2

0.00 50 100 150 200

Current density (mA.cm-2

)

Cell p

oten

tial (V

)

250 300 350 400 450

1.2

1.0

0.8

0.6

0.4

0.2

0.00 50 100 150 200

Current Density (mA.cm-2

)

Cell P

oten

tial (V

)

250 300 350 400 450

C2 : P 500 psi. t 2 min. T 160°C

C4 : P 500 psi. t 5 min. T 160°C

C5 : P 1500 psi. t 2 min. T 100°CC6 : P 1500 psi. t 5 min. T 100°C

C7 : P 1500 psi. t 2 min. T 160°CC8 : P 1500 psi. t 5 min. T 160°C

(a)

(b)

FIGURE 2.18 The effects of hot-pressing temperature and time on the polarization curves of

MEAs prepared at (a) 500 psi and (b) 1500 psi. [72]. (For color version of this figure, the reader is

referred to the online version of this book.)


2.4. FLOW FIELD DESIGN

Flow field plates or bipolar plates are key components of PEM fuel cells andstacks. In a PEM fuel cell, the functions of a flow field plate are as follows: (1) itprovides flow channels for the fuel and oxidant gases to their respective anodicand cathodic electrode surfaces, (2) it provides flow channels for the removal ofthe water coming from the humidifier and generated by electrochemical reac-tions, (3) it provides mechanical support for the anodic and cathodic electrodes,(4) it serves as a current collector, although a separate current collector is oftenused in a single cell, (5) it electronically connects one cell to another in a stack,and (6) it acts as a physical barrier to prevent the fuel, oxidant, and coolantfluids from mixing. In addition, a flow field plate is helpful for heat manage-ment. Flow field plates must perform the above functions simultaneously toachieve a good fuel cell performance. However, the requirements sometimesconflict, necessitating an optimized flow field design. In the past severaldecades of PEM fuel cell development, many flow field plate designs have beentested. Li et al. [86] extensively reviewed the subject. Figure 2.19 shows severaltypical flow field designs.

Pin-type flow field Straight and parallel flow field

Flow Channel

Flow-Channel

Rib or Channel Support

Inlet

Outlet

Serpentine flow field Interdigitated flow field

FIGURE 2.19 Several typical flow field designs [86].


2.4.1. Materials for a Flow Field Plate

Material selection is very important in designing a flow field plate, given thefunctions it performs in a PEM fuel cell. The materials must have a goodelectronic conductivity, low gas permeability, good mechanical stability tosupport the electrodes, good chemical/electrochemical stability, goodmachinability for making the flow field, light weight, and low cost. The mostcommonly used materials for flow field plates in PEM fuel cells are metalplates [87,88] and graphite plates [89], although other composite materials arealso used [90–92]. Metal plates have good electronic conductivity and can bemanufactured very thin to achieve light weights [88]. But their disadvantage isthat they get corroded in the PEM fuel cell operating environment, whichleads to flow field plate failure and thereby shortens the fuel cell’s lifetime.Gold coatings are often applied to metal flow field plates to solve this problem.At the present stage of PEM fuel cell technology, graphite plates are widelyused because they have good electronic conductivity and excellent chemical/electrochemical stability [93,94]. The disadvantages of graphite plates areobvious, though: high cost, brittleness, and greater weight compared to metalplates.

2.4.2. Flow Field Layout (Channel Pattern)

Flow field layout plays an important role in flow field design, as it affects bothreactant gas distribution and water removal. Several flow field layouts havebeen developed, according to their flow channels: the pin-type flow field[95,96], the straight parallel flow field [97,98], the serpentine flow field [98,99],and the interdigitated flow field [100–103], as shown in Fig. 2.19. All thesepatterns have their own characteristics in terms of reactant gas distribution,water, and heat management, and their advantages and disadvantages have beenextensively reviewed by Li and Sabir [86].

Of all the patterns, the straight parallel flow field and serpentine flow fieldare presently the mostly widely used; these patterns are shown in Figs 1.10 and1.11 of Chapter 1. In the straight parallel flow field, the gas distributions and thepressure along the flow channels are nonuniform due to the lack of change in theflow channel. In addition, the nonuniform pressure drops caused by the shortflow channel can result in low water removal ability, which has been addressedin Section 1.5.6 of Chapter 1. The serpentine flow field design can be classifiedas either single-channel or multichannel (the latter having two or moreserpentine channels). The single-channel serpentine flow field has a long flowchannel, leading to a large pressure drop along the flow channel between the gasinlet and outlet, which is good for water removal but results in nonuniform gasdistribution. The multichannel serpentine flow field has more flow channels, andmore reactant gas flows along the channels, leading to a uniform gas distributioninside the flow field plate. However, the drop in gas pressure between the inlet


and outlet is lower compared to the drop in the single-channel flow field. Thus,the water removal ability is relatively low [104].

2.4.3. Flow Channel Parameters

The channel geometry can significantly affect cell performance [105–107]because of its impact on the reactant gas flow and distribution, as well as thewater management inside the flow field. Figure 2.20 shows a cross-section offlow field channels. The main geometric parameters of flow channels are thechannel width w, channel depth d, rib/land width l, and wall angle q. Channellength L is another geometric parameter that is dependent on the sizeconstraint of the particular application. Given a certain size and layout fora flow field, optimization of the geometric parameters of the flow channelscan yield a more uniform reactant gas distribution as well as better water andheat management, leading to better fuel cell performance. The typicalparameters for the flow channels are 0.5–2.5 mm for the channel width,0.2–2.5 mm for the channel depth, 0.2–2.5 mm for the rib width, and 0–15�for the wall angle [99].

2.4.4. Flow Field Plate Fabrication

Flow channels can be made on either or both sides of a graphite, metal, orcomposite material plate. In a fuel cell stack, a plate with flow channels on bothsides is called a bipolar plate. The fabrication process of a flow field platedepends on the materials used. Carbon-based plates (using materials such asgraphite felt, flexible graphite, and carbon resin) are usually brittle; conse-quently, the plate thickness and the cross-sectional flow channel area should bemade larger during fabrication. Metal plates usually have excellent mechanical

FIGURE 2.20 Cross-section of flow field channels: rib width l, channel width w, channel depth

d, and wall angle q [99,107]. (For color version of this figure, the reader is referred to the online

version of this book.)


stability and can be made into thinner flow field plates with a smaller cross-sectional area. Although metal plates usually have good electronic conduc-tivity, they are subject to corrosion in a PEM fuel cell operating environment,resulting in increased surface resistance and decreased PEM fuel cell perfor-mance. Therefore, protective coatings are usually applied to metal plates toprevent flow field plate corrosion [87,88].

Flow field fabrication is a purely mechanical process; therefore, anyworkshop with appropriately designed machines can accomplish it.

2.5. SEALING DESIGN

The sealing gasket is yet another important component in both single PEM fuelcells (as shown in Fig. 1.5 in Chapter 1) and stacks (as shown in Fig. 2.12). Twosealing gaskets are required for one MEA unit, placed between each side of theMEA and the flow field plates. Generally, the sealing gaskets perform threeimportant functions in a PEM fuel cell or stack: (1) sealing off gases to preventgas crossover and leakage, (2) insulating to prevent the fuel cell from shorting,and (3) sealing the coolant. Sealing gaskets must be designed carefully becausea poor-quality gasket can cause fuel cell leakage, leading to low performanceand safety issues, especially when the fuel is H2 gas.

2.5.1. Sealing Material Selection

To perform their functions, sealing materials in PEM fuel cells must meetstringent criteria: (1) electronic insulation, (2) high chemical/electrochemicalstability, (3) suitable compressibility, and (4) good compatibility with thereactant gases and coolant fluids. The most widely used sealing materials arePTFE, fluoroelastomer, and silicon-based materials such as silicon rubber andsilicon elastomers [108–110]. If the sealing materials are not stable enoughafter long-term PEM fuel cell operation, the decomposed impurities or otherproducts may get into the fuel cell components [108–111], leading tocontamination. For example, some decomposition products of silicon-basedmaterials may get adsorbed on the membrane and/or CL, causing decreasedmembrane conductivity and/or catalyst activity, and certain decompositionfragments may get into the GDL, changing its hydrophilicity/hydrophobicity.

2.5.2. Sealing Design and Fabrication

In accordance with MEA and flow field designs, many sealing designs havebeen developed during the past several decades [108,112]. The major materialsused are silicon rubbers. Normally, two methods are used to fabricate thesesealings: die cutting or molding, such as screen printing. In some fabricationmethods, the sealings are directly molded onto the GDL [100], bipolar plate[101,102], and MEA, as shown in Fig. 2.21 [103–105]. In recent years, the

FIGURE 2.21 Schematic representation of a seal integrated with an MEA [112]. (For color

version of this figure, the reader is referred to the online version of this book.)


design shown in Fig. 2.21 has become more popular with CCM development.The CCM is usually laminated together with one sealing gasket sheet on eachside to form a sealed MEA. Another popular sealing arrangement is simply toput a gasket sheet between the flow field plate and the MEA, and then pressthem all together during fuel cell assembly, as shown in Fig. 1.5 in Chapter 1.

2.6. SINGLE CELL DESIGN AND ASSEMBLY

2.6.1. Single Cell Hardware

Single fuel cell hardware, shown in Fig. 1.5 of Chapter 1, consists of end plates,current collectors, sealing gaskets, flow field plates, and bolts. Depending onthe materials used, single cell hardware design should consider several factors,including but not limited to material selection, stability, flow field layout, andflow channel pattern. However, the cell hardware components must be designedto achieve optimal combined performance, including compatibility with theMEA.

2.6.2. Single Cell Assembly

Single cell assembly is the process of putting together all the requisitecomponents to form a single fuel cell. As shown in Fig. 1.5 in Chapter 1, all thecomponents are pressed together with the MEA in the middle, then tightenedusing bolts. In some tests, to exert a uniform and constant pressure along theMEA surface, a gas bladder is used with a piston along one side of the singlecell. The pressure used to hold the components together is important inachieving a high-performance single cell. For example, if the pressure is toohigh, the porosity of the GDE may be reduced; in the worst-case scenario, theMEA may be damaged, leading to a large mass transfer resistance andconsequently poor fuel cell performance. However, if the pressure is too low,the contacts between the fuel cell components will be an issue, again resultingin large internal resistance and consequently low cell performance. In our

FIGURE 2.22 A designed single PEM fuel cell with an active area of 4.4 cm2 [113]. (For color

version of this figure, the reader is referred to the online version of this book.)


laboratory, the designed fuel cell hardware includes bladder plates, piston,plastic plates, current collectors, flow field plates, and gaskets, as shown inFig. 2.22 [113]. Note that on the end plates there are the gas inlet and outletmouths to connect to the gas manifolds, for fuel and oxidant feeding, which arenot shown in this figure. The plastic plate is used to isolate the current collector/flow field plate from the metal bladder. Gaskets are attached to the flow fieldplates to seal the MEA on both sides. The cell is pressed and sealed using thebladder plates and piston, powered by gas (e.g. nitrogen or air). The tightness ofthe fuel cell can easily be controlled by adjusting the bladder pressure.

2.7. STACK DESIGN AND ASSEMBLY

2.7.1. Hardware of a Fuel Cell Stack

As shown in Fig. 2.23, a fuel cell stack consists of many single cells connectedin series. It also contains end plates, flow field plates (current collector),gaskets, and bolts. The difference is that there is a cooling plate between eachtwo adjacent single cells. Due to the heat generated during fuel cell operation,the extra heat must be removed to maintain the operating temperature.Therefore, this cooling plate is necessary. Another method is to integrate theanode (or cathode) flow field plate of one cell, the cooling plate, and thecathode (or anode) flow field plate of the adjacent cell into one shared plate. On

FIGURE 2.23 Hardware of an H2/air PEM fuel cell stack [61].


this shared plate, both sides are fabricated with flow field channels, then oneside serves as the anode (or cathode) flow field of the one cell and the other asthe cathode (or anode) flow field of the other single cell. Inside this bipolar plateare some channels for coolant flow.

2.7.2. Heat Management in a PEM Fuel Cell Stack

Normally, PEM fuel cells are operated between 60 and 80 �C. At the startup ofa fuel cell stack, the electrochemical reactions inside the stack produce energythat will rapidly heat the whole stack to this temperature range within 1–2 min.Due to the limitations of the PEM (e.g. Nafion� membrane), the stacktemperature must be kept <95 �C, otherwise the fuel cell performance willrapidly decline. Of course, if another high-temperature membrane is used, thestack operating temperature can be elevated. To maintain the stack temperaturewithin a desired range, such as 60–80 �C, a stack cooling system must be addedto the fuel cell system to remove the extra heat. It is also well recognized thatthe temperature distribution in a stack affects the fuel cell’s performance. Thecooling system is therefore helpful in homogenizing the temperature distri-bution within the stack.

The cooling system in a fuel cell system can be designed as either internalcooling [114] or external cooling [115]. Within these categories, internalcooling can be either air or liquid cooling (liquid being water or a waterethylene glycol mixture). Normally, an air or liquid cooling plate can bedesigned to be integrated with the bipolar plate, to simplify the fuel cell stack[114,116]. The coolant plates in a fuel cell stack must be sealed to prevent fluidfrom leaking into the reactant gas channels or outside of the fuel cell. For theexternal cooling system, there is no cooling liquid present in the cell active


area, which eliminates any sealing problems with respective to the electrode[115] and yields a simplified fuel cell system.

2.7.3. Fuel Cell Stack Assembly

A fuel cell stack is assembled by packing many single cells in series, as shownin Fig. 2.23. The electronic series connection of all these single cells is realizedby the electronic conducting bipolar plates. The number of single cells dependson the desired stack power and size, and the performance of single cells. Inother words, the power that can be generated by a fuel cell stack is determinedby the number of its cells, the total active area of the MEA, and the single cellperformance.

2.8. CHAPTER SUMMARY

This chapter has presented the design and assembly of single PEM fuel cellsand stacks, with a focus on the MEA and the fabrication of its components. Thisis because the MEA is the most important component of a PEM fuel cell, whereelectrochemical reactions occur for power generation. The chapter discussesMEA design and assembly/fabrication, including GDLs (gas diffusion mediumand microporous sublayer), CLs, and PEMs, as well as analyzes various factorsthat affect MEA design and performance. The chapter also provides a typicalexample of step-by-step MEA fabrication from various component materials tothe whole MEA. Other components, including flow field plates, sealing gaskets,and their corresponding designs and fabrication, are also presented in thischapter. Finally, the design and fabrication of both single fuel cells and stacksare introduced, the intent being to provide readers with the basic informationand procedures for achieving workable PEM fuel cell and stack hardware. Webelieve that this chapter forms a solid foundation for the PEM fuel cell testingand diagnosis described in the following chapters.

REFERENCES

[1] Cindrella L, Kannan AM, Lin JF, Saminathan K, Ho Y, Lin CW, et al. J Power Sources

2009;194:146–60.

[2] Du C, Wang B, Cheng X. J Power Sources 2009;187:505–8.

[3] Kannan AM, Munukutla L. J Power Sources 2007;167:330–5.

[4] Lin JF, Wertz J, Ahmad R, Thommes M, Kannan AM. Electrochim Acta 2010;55:2746–51.

[5] Fuel Cells Bull 2001;4. 15–15.

[6] Ko T-H, Liao Y-K, Liu C-H. New Carbon Mater 2007;22:97–101.

[7] Ko T-H, Liao Y-K, Liu C-H. Carbon 2007;45. 2321–2321.

[8] Liu C-H, Ko T-H, Kuo W-S, Chou H-K, Chang H-W, Liao Y-K. J Power Sources

2009;186:450–4.

[9] Liu C-H, Ko T-H, Shen J-W, Chang S-I, Chang S-I, Liao Y-K. J Power Sources

2009;191:489–94.


[10] Yang H, Tu HC. Chiang IL. Int J Hydrogen Energy 2010;35:2791–5.

[11] Zhang F-Y, Advani SG, Prasad AK. J Power Sources 2008;176:293–8.

[12] Mathias M, Roth J, Fleming J, Lehner W. In: Vielstich W, Gasteiger HA, Lamm A, editors.

Handbook of fuel cells-fundamentals, technology and applications, vol. 3. John Wiley &

Sons, Ltd.; 2003.

[13] Lim C, Wang CY. Electrochim Acta 2004;49:4149–56.

[14] Pai Y-H, Ke J-H, Huang H-F, Lee C-M, Zen J-M, Shieu F-S. J Power Sources

2006;161:275–81.

[15] Park G-G, Sohn Y-J, Yang T-H, Yoon Y-G, Lee W-Y, Kim C-S. J Power Sources

2004;131:182–7.

[16] Jordan LR, Shukla AK, Behrsing T, Avery NR, Muddle BC, Forsyth M. J Power Sources

2000;86:250–4.

[17] Tseng C-J, Lo S-K. Energy Convers Manage 2010;51:677–84.

[18] Park S, Lee J-W, Popov BN. J Power Sources 2006;163:357–63.

[19] Kong CS, Kim D-Y, Lee H-K, Shul Y-G, Lee T-H. J Power Sources 2002;108:185–91.

[20] Qi Z, Kaufman A. J Power Sources 2002;109:38–46.

[21] Nam JH, Kaviany M. Int J Heat Mass Transf 2003;46:4595–611.

[22] Wang X, Zhang H, Zhang J, Xu H, Zhu X, Chen J, et al. J Power Sources 2006;162:474–9.

[23] Passalacqua E, Squadrito G, Lufrano F, Patti A, Giorgi L. J Appl Electrochem 2001;31:

449–54.

[24] Lee H-K, Park J-H, Kim D-Y, Lee T-H. J Power Sources 2004;131:200–6.

[25] Ambrosio E, Francia C, Gerbaldi C, Penazzi N, Spinelli P, Manzoli M, et al. J Appl

Electrochem 2008;38:1019–27.

[26] Moreira J, Ocampo AL, Sebastian PJ, Smit MA, Salazar MD, del Angel P, et al. Int J

Hydrogen Energy 2003;28:625–7.

[27] Park S, Lee J-W, Popov BN. J Power Sources 2008;177:457–63.

[28] Giorgi L, Antolini E, Pozio A, Passalacqua E. Electrochim Acta 1998;43:3675–80.

[29] Antolini E, Passos RR, Ticianelli EA. J Power Sources 2002;109:477–82.

[30] Wang XL, Zhang HM, Zhang JL, Xu HF, Tian ZQ, Chen J, et al. Electrochim Acta

2006;51:4909–15.

[31] Paganin VA, Ticianelli EA, Gonzalez ER. J Appl Electrochem 1996;26:297–304.

[32] Lufrano F, Passalacqua E, Squadrito G, Patti A, Giorgi L. JAppl Electrochem 1999;29:445–8.

[33] Antolini E, Passos RR, Ticianelli EA. J Appl Electrochem 2002;32:383–8.

[34] Neergat M, Shukla AK. J Power Sources 2002;104:289–94.

[35] Wang X, Hsing IM, Yue PL. J Power Sources 2001;96:282–7.

[36] Zhang H, Wang X, Zhang J, Zhang J. In: Zhang J, editor. PEM fuel cell electrocatalysts and

catalyst layersdfundamentals and applications. London: Springer; 2008. p. 889–916.

[37] Wilson MS, Valerio JA, Gottesfeld S. Electrochim Acta 1995;40:355–63.

[38] Raistrick ID. US Patent 4876115, 1989.

[39] Wilson MS, Gottesfeld S. J Electrochem Soc 1992;139:L28–30.

[40] Springer TE, Wilson MS, Gottesfeld S. J Electrochem Soc 1993;140:3513–26.

[41] Wilson MS, Gottesfeld S. J Appl Electrochem 1992;22:1–7.

[42] Wilson MS. Membrane catalyst layer for fuel cells. In: US Patent 5234777, 1993.

[43] O’Hayre R, Lee S-J, Cha S-W, Prinz FB. J Power Sources 2002;109:483–93.

[44] Zhang J, Zhang J. In: Zhang J, editor. PEM fuel cell electrocatalysts and catalyst

layersdfundamentals and applications. London: Springer; 2008. p. 965–1002.

[45] Srinivasan S, Velev OA, Parthasarathy A, Manko DJ, Appleby AJ. J Power Sources

1991;36:299–320.


[46] Shin SJ, Lee JK, Ha HY, Hong SA, Chun HS, Oh IH. J Power Sources 2002;106:146–52.


[48] Dhar HP. Method for catalyzing a gas diffusion electrode. In: US Patent 5521020, 1996.




[52] He C, Qi Z, Hollett M, Kaufman A. Electrochem Solid State Lett 2002;5:A181–3.

[53] Song SQ, Liang ZX, Zhou WJ, Sun GQ, Xin Q, Stergiopoulos V, et al. J Power Sources

2005;145:495–501.

[54] Xie J, More KL, Zawodzinski TA, Smith WH. J Electrochem Soc 2004;151:A1841–6.

[55] Saha MS, Paul DK, Peppley BA, Karan K. Electrochem Commun 2010;12:410–3.

[56] Zhang J, Wang X, Hu J, Yi B, Zhang H. Bull Chem Soc Jpn 2004;77:2289–90.

[57] Hsu CH, Wan CC. J Power Sources 2003;115:268–73.

[58] Mussell RD. Active layer for membrane electrode assembly. In: US Patent 5882810 1999.

[59] Debe MK. In: Vielstich W, Gastieger HA, Lamm A, editors. Handbook of fuel cells-

fundamentals, technology and applications. John Wiley & Sons; 2003. p. 576–89.

[60] Song Y, Fenton JM, Kunz HR, Bonville LJ, Williams MV. J Electrochem Soc

2005;152:A539–44.

[61] Mehta V, Cooper JS. J Power Sources 2003;114:32–53.

[62] Alberti G, Casciola M. Ann Rev Mater Res 2003;33:129–54.

[63] Brandon NP, Skinner S, Steele BCH. Ann Rev Mater Res. 2003;33:183–213.

[64] Schuster MFH, Meyer WH. Ann Rev Mater Res 2003;33:233–61.

[65] Paddison SJ. Ann Rev Mater Res 2003;33:289–319.

[66] Roziere J, Jones DJ. Ann Rev Mater Res 2003;33:503–55.

[67] Nakao M, Yoshitake M. In: Vielstich W, Gasteiger HA, Lamm A, editors. Handbook of fuel

cells-fundamentals, technology and applications. John Wiley & Sons, Ltd; 2003. p. 412–9.

[68] Jones DJ, Roziere J. In: Vielstich W, Gasteiger HA, Lamm A, editors. Handbook of fuel

cells-fundamentals, technology and applications. John Wiley & Sons, Ltd; 2003.

[69] Hickner MA, Ghassemi H, Kim YS, Einsla BR, McGrath JE. Chem Rev 2004;104:

4587–612.

[70] Lin J-C, Lai C-M, Ting F-P, Chyou S-D, Hsueh K-L. J Appl Electrochem 2009;39:1067–73.

[71] Liang ZX, Zhao TS, Xu C, Xu JB. Electrochim Acta 2007;53:894–902.

[72] Therdthianwong A, Manomayidthikarn P, Therdthianwong S. Energy 2007;32:2401–11.

[73] Zhang J, Yin G-P, Wang Z-B, Lai Q-Z, Cai K-D. J Power Sources 2007;165:73–81.

[74] Liu P, Yin G-P, Wang E-D, Zhang J, Wang Z-B. J Appl Electrochem 2009;39:859–66.

[75] Yan W-M, Wu D-K, Wang X-D, Ong A-L, Lee D-J, Su A. J Power Sources

2010;195:5731–4.

[76] Ticianelli EA, Derouin CR, Redondo A, Srinivasan S. J Electrochem Soc 1988;135:

2209–14.

[77] Srinivasan S, Ticianelli EA, Derouin CR, Redondo A. J Power Sources 1988;22:359–75.

[78] Kumar GS, Raja M, Parthasarathy S. Electrochim Acta 1995;40:285–90.

[79] Prasanna M, Cho EA, Lim TH, Oh IH. Electrochim Acta 2008;53:5434–41.

[80] Cha SY, Lee WM. J Electrochem Soc 1999;146:4055–60.

[81] Suzuki T, Tsushima S, Hirai S. Int J Hydrogen Energy 2011;36:12361–9.

[82] Jeon S, Lee J, Rios GM, Kim H-J, Lee S-Y, Cho E, et al. Int J Hydrogen Energy

2010;35:9678–86.

[83] Cho HJ, Jang H, Lim S, Cho E, Lim T-H, Oh I-H, et al. Int J Hydrogen Energy

2011;36:12465–73.


[84] Kim KH, Kim HJ, Lee KY, Jang JH, Lee SY, Cho E, et al. Int J Hydrogen Energy

2008;33:2783–9.

[85] Kim K-H, Lee K-Y, Kim H-J, Cho E, Lee S-Y, Lim T-H, et al. Int J Hydrogen Energy

2010;35:2119–26.

[86] Li X, Sabir I. Int J Hydrogen Energy 2005;30:359–71.

[87] Wind J, LaCroix A, Braeuninger S, Hedrich P, Heller C, Schudy M. In: Vielstich W,

Gasteiger HA, Lamm A, editors. Handbook of fuel cells-fundamentals, technology and

applications. John Wiley & Sons, Ltd; 2003. p. 294–307.

[88] Tawfik H, Hung Y, Mahajan D. J Power Sources 2007;163:755–67.

[89] Robberg K, Trapp V. In: Vielstich W, Gasteiger HA, Lamm A, editors. Handbook of fuel

cells-fundamentals, technology and applications. John Wiley & Sons, Ltd; 2003. p. 308–14.

[90] Dweiri R, Sahari J. J Power Sources 2007;171:424–32.

[91] Maheshwari PH, Mathur RB, Dhami TL. J Power Sources 2007;173:394–403.

[92] Mercuri RA, Gough JJ. US Patent 6037074, 2000.

[93] Qian P, Zhang H, Chen J, Wen Y, Luo Q, Liu Z, et al. J Power Sources 2008;175:613–20.

[94] Martin J, Oshkai P, Djilali N. J Fuel Cell Sci Technol 2005;2:70–80.

[95] Reiser CA, Sawyer RD. US Patent 4769297, 1988.

[96] Reiser CA. US Patent 4826742, 1989.

[97] Voss HH, Chow CY. US Patent 5230966A, 1993.

[98] Su A, Chiu YC, Weng FB. Int J Energy Res 2005;29:409–25.

[99] Wilkinson DP, Vanderleeden O. In: Vielstich W, Gasteiger HA, Lamm A, editors. Handbook

of fuel cell-fundamentals, technology and applications. John Wiley & Sons, Ltd; 2003.

[100] Nguyen TV, He E. In: Vielstich W, Gasteiger HA, Lamm A, editors. Handbook of fuel cell-

fundamentals, technology and applications. John Wiley & Sons, Ltd; 2003.

[101] Guo SM. In: Sunderam V, Albada G, Sloot P, Dongarra J, editors. Computational

sciencedICCS 2005. Springer Berlin/Heidelberg; 2005. p. 104–11.

[102] Yamada H, Hatanaka T, Murata H, Morimoto Y. J Electrochem Soc 2006;153:A1748–54.

[103] Wilson MS. US Patent 005641586A, 1997.

[104] Zhang J, Li H, Shi Z, Zhang J. Int J Green Energy 2010;7:461–74.

[105] Wang X-D, Yan W-M, Duan Y-Y, Weng F-B, Jung G-B, Lee C-Y. Energy Convers Manage

2010;51:959–68.

[106] Shimpalee S, Van Zee JW. Int J Hydrogen Energy 2007;32:842–56.

[107] Akhtar N, Qureshi A, Scholta J, Hartnig C, Messerschmidt M, Lehnert W. Int J Hydrogen

Energy 2009;34:3104–11.

[108] Frisch L. Sealing Technol 2001;2001:7–9.

[109] Schulze M, Knori T, Schneider A, Gulzow E. J Power Sources 2004;127:222–9.

[110] Tan J, Chao YJ, Yang M, Lee W-K. Van Zee JW. Int J Hydrogen Energy 2011;36:1846–51.

[111] Du B, Guo Q, Pollard R, Rodriguez D, Smith C, Elter J. J Oral Microbiol 2006;58:45–9.

[112] Koch S. PEM fuel cell stack sealing, report 2005 (Powerpoint slides; available online http://

www.slideserve.com/Mercy/pem-fuel-cell-stack-sealing).

[113] Tang Y, Zhang J, Song C, Liu H, Zhang J, Wang H, et al. J Electrochem Soc

2006;153:A2036–43.

[114] Wu J, Galli S, Lagana I, Pozio A, Monteleone G, Yuan XZ, et al. J Power Sources

2009;188:199–204.

[115] Scholta J, Messerschmidt M, Jorissen L, Hartnig C. J Power Sources 2009;190:83–5.

[116] Peng J, Shin J-Y, Lee S-J. US Patent 20090162713A1, 2009.

http://www.slideserve.com/Mercy/pem-fuel-cell-stack-sealing

http://www.slideserve.com/Mercy/pem-fuel-cell-stack-sealing

pem fuel cell testing and diagnosis || design and fabrication of pem fuel cell mea, single cell, and...

Documents